Hey, have you guys heard about OpenAI's new o1 model? They dropped it yesterday, and it’s… interesting, but maybe not as game-changing as everyone hoped. So, it’s called o1—codenamed “Strawberry”—and the big thing about it is that it “pauses to think” before giving answers, especially for more complex problems. Sounds great, right? But here’s the catch: it’s crazy expensive and kind of slow compared to the previous GPT-4o model.
It’s supposed to be really good at reasoning through big questions, breaking them into smaller steps and all that. In fact, if you’ve got some complicated thing to figure out—like, planning a huge Thanksgiving dinner for 11 people with just two ovens—o1 might be perfect. It thinks through all the angles, even suggesting renting a portable oven (who knew that was even a thing?!). But, when it comes to simple questions, like where to find cedar trees in America, it totally overthinks it—giving 800+ words where Gpt-40
